Jun 02 2010

Potential New Mechanism of Pain Relief Discovered

The development of drugs and other treatments for specific symptoms or conditions relies heavily on either serendipity (the chance finding of a beneficial effect) or on an understanding of underlying mechanisms. In pain, for example, there are limited ways in which we can block pain signals – such as activating opiate receptors, or inhibiting prostaglandins. There are only so many ways in which you can interact with these systems. The discovery of a novel mechanism of modulating pain is therefore most welcome, and has the potential of leading to entirely new treatments that may have a better side effect profile than existing treatments and also have an additive clinical effect.

A recent study by Nana Goldman et. al., published in Nature Neuroscience, adds to our understanding of pain relief by identifying the role of adenosine in reducing pain activity in the peripheral nervous system. The researchers, in a nice series of experiments, demonstrated that producing a local painful stimulus in mice causes the local release of ATP (adenosine triphosphate) that peaks at about 30 minutes. This correlates with a decreased pain response in the mice. Further, if drugs are given that prolong the effect of adenosine, the analgesic effect itself is prolonged.

Also, if drugs are given that activate the adenosine A1 receptor, the observed analgesic effect is replicated. When these experiments are replicated in knockout mice that do not have the gene for the adenosine A1 receptor, there is no observed analgesic effect.

Together these experiments are fairly solid evidence that local pain results in the local release of adenosine that in turn binds to the adenosine A1 receptor inhibiting the pain response. This is potentially very exiting – it should lead to further investigation of the adenosine A1 receptor and the effects of activating and inhibiting it. This may lead to the development of drugs or other interventions that activate these receptors and may ultimately be a very useful addition to our ability to treat acute and chronic pain.

Now Comes the Spin

In a rational science-based world the above would be the end of this blog entry. But that is not the world we currently live in. In this study the chosen method of provoking pain was the insertion of an acupuncture needle into the “Zusanli point” of the lower extremity and rotating it. This allowed the authors of this study to spin the results as if they were validation for acupuncture itself – and the mainstream press dutifully followed suit.

The introduction to this study is an apology for acupuncture itself, and sets the tone for the rest of the paper as well as the press release and comments to follow:

Acupuncture is a procedure in which fine needles are inserted into an individual at discrete points and then manipulated, with the intent of relieving pain. Since its development in China around 2,000 B.C., acupuncture has become worldwide in its practice. Although Western medicine has treated acupuncture with considerable skepticism, a broader worldwide population has granted it acceptance. For instance, the World Health Organization endorses acupuncture for at least two dozen conditions and the US National Institutes of Health issued a consensus statement proposing acupuncture as a therapeutic intervention for complementary medicine. Perhaps most tellingly, the U.S. Internal Revenue Service approved acupuncture as a deductible medical expense in 1973.

That this paragraphs appears in a high impact peer-reviewed journal is very curious (to put it mildly). It is full of misinformation and logical fallacies. Acupuncture is used for more than pain relief, and so making statements about the efficacy of “acupuncture” go way beyond the scope of this study, which only involves pain. The authors should have specifically pointed out that this study cannot be used to explain any indication for acupuncture that does not involve local pain relief. Acupuncture as it was practiced in China 4000 years ago bears little resemblance to what is practiced today, and may have been more of a form a bloodletting.

The authors then make a false dichotomy between “Western medicine” and, presumably, “Eastern medicine” – which is an anti-scientific and culturally bigoted point of view. This is followed by an argument from popularity, and a particularly bizarre argument from authority – noting that the IRS accepts acupuncture as a medical expense. I was unaware that the IRS is a scientific medical organization.

The authors also fail to put this study into its proper context by reviewing existing evidence – the more traditional use of the introduction to peer-reviewed research. The clinical research on acupuncture, which is quite extensive, finds that any measured symptomatic effect is almost certainly due mostly or entirely to placebo effects. Further, it does not matter where you stick the acupuncture needle, nor even if you do stick the need through the skin. Twisting toothpicks against the skin in random locations has the same effect.

My interpretation of the research is that acupuncture (placing needles at specific acupuncture points to manipulate chi) does not work. There may, however, be non-specific physiological responses to the mechanical stimulation of sticking needles at random locations, or just poking toothpicks. This study, if anything, supports this interpretation – it, if fact, has identified a local mechanism of analgesia that can help explain a non-specific response to acupuncture, sham acupuncture, or placebo acupuncture and therefore is consistent with the clinical evidence showing no difference among these interventions.

This is also not the first study to identify potential local mechanisms of pain relief from mechanical stimulation. A 2002 study correlated insertional activity (depolarization of muscle fibers in response to needle insertion) or electrical stimulation with pain relief. This would explain why, in this study, constant twisting of the needle was needed to provoke pain relief – perhaps insertional muscle activity is needed to release adenosine, or increases its release.

By focusing on what is really going on here we can best understand how to develop methods to optimally capitalize on these local mechanisms for pain relief. It must be noted, however, that needle insertion gives very unimpressive results in clinical trials. It may be that the effect is too temporary to be worthwhile, at least as mechanically provoked – pharmacologically activating the adenosine A1 receptor may be a better strategy. Also, it is very difficult to extrapolate from mouse data as they are much smaller than humans, and therefore their nerves and motor end-plates (the location where the nerves innervate the muscles) are much closer together and superficial – closer to the skin.

Mechanisms are interesting, but net clinical outcomes in humans are the only kind of scientific data that really tells us if a modality works or not.


Finally, it has to be emphasized that this study says nothing about acupuncture itself, except for providing a possible mechanism for a non-specific local response. The term “acupuncture”, in fact, is becoming increasingly problematic and is confusing the scientific literature, not to mention the public. What is acupuncture? If we use the term broadly enough to mean any use of needles, with or without electrical stimulation, at any points, with or without skin penetration, etc. then the term is too broad to be useful. If we use the term narrowly – to mean sticking needles to a certain depth in specific acupuncture points that work through a novel mechanism specific to those locations, then we can say, based upon extensive research, that “acupuncture” does not work and its proposed underlying mechanisms are nothing more than pre-scientific superstition.

This study is an excellent example of the mischief caused by confusing the non-specific use of the term “acupuncture” with its more traditional use. Research involving acupuncture in its vaguest sense is used to promote “acupuncture” in the traditional sense. This is highly deceptive and scientifically sloppy.

The researchers of this current study could have used other controls to see if the effect they discovered is in any way specific to any acupuncture variables. For example – they could have used a non-acupuncture point as a control, or other forms of mechanical pain production that do not involve needles. I suspect any local pain production or mechanical trauma beyond a certain threshold would result in the same adenosine response – which certainly seems like a non-specific mechanism to modulate pain.

Because they did not do this they did not actually research “acupuncture”. The description of this research in the published paper and in the press should have been as I discussed in the opening of this post. Instead genuinely interesting research that may lead to novel pain treatments is being diverted as propaganda for an ancient superstition.


Note: This article is cross-posted over at Science-Based Medicine


60 responses so far

60 Responses to “Potential New Mechanism of Pain Relief Discovered”

  1. SARAon 02 Jun 2010 at 9:31 am

    I’m not a scientist, so please tell me where I missed something.
    On the one hand you point out that there is no evidence that acupuncture (poking needles at chi points) works beyond placebo. But then you cite studies that show needles or even tooth picks produce some non specific pain relief. And if I understand the experiment, this might be due to the adenosine.

    Even as a layman I thought their introduction was full of illogical implications – in particular I thought it odd that the IRS is used to support medical efficacy.

    So if I understand you – acupuncture is considered not effective because it doesn’t consistently predict a specific response to a specific mechanism/location. However, there is reason to suppose that stimulation to create a local pain does cause non-specific pain relief, but not predictably.

    For something to be considered medical treatment, it should be a specific procedure leading to predictable results.

  2. Steven Novellaon 02 Jun 2010 at 9:41 am

    The main problem is the definition of acupuncture. Scientific terms need to be specific to be useful – hopefully they are “operational”, meaning there are specific features that must be present and others that must be absent.

    So – what is acupuncture? Research has shown that anything which can be said to be specific to acupuncture has no effect.

    Poking people in random locations, with or without penetration, however, has non-specific psychological and physiological effects. This is not “acupuncture” though, any more than some at-home stretching is chiropractic.

    What I think we can say at this point is:

    1 – Acupuncture is not based on scientific principles and does not work for anything.
    2 – Low levels of mechanical trauma provoke local tissue and neurological responses that inhibit local pain and inflammation – not surprising at all, and entirely incidental to the application of “acupuncture.”
    3 – These effects are temporary and short term, and probably local, and in clinical trials are of questionable significance and utility.
    4 – However, understanding the underlying mechanisms may lead to more useful therapies with more convenient application and longer duration of effects. But we will need to go through the entire development process to see how these pan out.

  3. Michael Meadonon 02 Jun 2010 at 9:48 am

    Superb write-up Steve. Posts like these is why I read your blog, and always hope you’ll write about stuff I send you. To gush a bit more, excellent work on being as neutral as possible. I haven’t seen anyone else (not that I’ve read a lot about this) praise the study’s discovery adenosine pain inhibition.

    Also: are you going to write a formal response for Nature Neuroscience? That might be useful.

  4. Michael Meadonon 02 Jun 2010 at 9:54 am

    Discovery OF…

  5. Eric Thomsonon 02 Jun 2010 at 10:06 am

    I second Michael’s suggestion that you write a letter to Nature Neuroscience on this. It would be interesting to see the author /journal reaction.

  6. Steven Novellaon 02 Jun 2010 at 10:12 am

    Two other excellent treatments of this article are:

    Orac at Respectful Insolence: http://scienceblogs.com/insolence/2010/06/when_what_an_acupuncture_study_shows_is.php

    And Ed Yong at Not Exactly Rocket Science: http://blogs.discovermagazine.com/notrocketscience/2010/05/30/a-biological-basis-for-acupuncture-or-more-evidence-for-a-placebo-effect/

    Your suggestion about writing a comment to Nature Neuroscience is a good one.

  7. ccbowerson 02 Jun 2010 at 11:56 am

    Did they have Mercola write that intro or did they take it from some new-age website? I’m amazed that the study was allowed to be published with that terrible introduction included. The logical fallacies are all over the place, but even putting that aside the intro has little/nothing to do with the actual study.

  8. Eric Thomsonon 02 Jun 2010 at 12:54 pm

    I just briefly read over the paper. As a naive reader (I know nothing about apuncture or studies of acupuncture) I came away with:
    1. Acupuncture is known to have analgesic properties (background of paper).
    2. Acupuncture needles inserted near the ‘Zusanli point’ (basically a location place near the knee) and rotated supresses pain in rats. If the rotation isn’t done, you don’t have the analgesic effects.
    3. Acupuncture needle insertion causes adenosine release, with much greater release with needle rotation.
    4. This pain supression is A1-receptor dependent (that’s an adenosine receptor).
    5. Introduction of an A1-receptor agonist (activator) is sufficient to reduce pain.
    1-5 suggest that the well-established analgesic effects of acupuncture are caused largely by adenosine release.

    This all seems actually fairly reasonable. Two main “skeptical” questions that Steven brought up I address below.

    1. Does the location matter?
    The authors didn’t address whether this would work anywhere on the body. That is, are there specific “centers” on the body where acupuncture would be particularly effective? Their study doesn’t address this.

    My hunch is that there may actually be localized regions where it is more effective. Namely, locations of high concentration of A1-receptors that are part of the ascending sensory pathways. My hunch is the adenosine release is quite general, but the effects of adenosine will likely be anisotropic. This at least suggests further research, e.g., labelling of peripheral A1-receptors.

    2. What about other stimuli besides needles?
    Novella brought this up at the end. On page 5 the authors admit things like massage might also cause ATP release/adenosine buildup. They think this might be plausible, and mention vibratory stimuli presented to the skin supress pain by release of adenosine. But they claim acupuncture is different because it is “typically applied to deep tissue, including muscle and connective tissue, and acupoints may better overlap with their proximity to ascending nerve tracks than to the density of cutaneous afferents.”

    This is basically what I suggested above.

    My response to this paper is that it is decent. They seem to have a pretty relaxed attitude, not particularly dogmatic. Needles cause release of adenosine, which supresses pain via A1-receptors. This explains acupuncture’s analgesic effects. It doesn’t seem particularly ludicrous to me.

    The suggestion that they should have left out the acupuncture angle of the paper seems proposterous to me. That was the main motivation, the method used, everything hinged on this being a study of acupuncture.

    Note, this is the response of a naive reader, I have done no investigations of acupuncture. There is a chance I have been mislead, but frankly I understand why this is in Nature Neuro, and frankly it doesn’t come off as new-agey or methodologically flawed. Yes, they left questions open, but no major flaws.

  9. sonicon 02 Jun 2010 at 2:21 pm

    The first sentence of the Introduction (quoted in the post)

    “Acupuncture is a procedure in which fine needles are inserted into an individual at discrete points and then manipulated, with the intent of relieving pain.”

    Any complaints about how poorly the term ‘acupuncture’ is defined do not belong with this article or its authors.

    The mention of the IRS is to point out
    “a broader worldwide population has granted it acceptance.”
    To call this mention an argument from authority and to note that the IRS isn’t a scientific medical organization is nonsequitur.

  10. Yousaf Parvezon 02 Jun 2010 at 2:37 pm

    It is a sad and regretful state of affairs when genuine scientific research is hijacked for unscrupulous commercial entities. This is also potentially very dangerous as the general public will be further duped in to spending vast sums of money annually towards a near useless “specialist” pain relieval method.
    Is this another case of sloppy journalism or is it a case of deliberate misguiding?

  11. bindleon 02 Jun 2010 at 2:49 pm

    “To call this mention an argument from authority and to note that the IRS isn’t a scientific medical organization is nonsequitur.”
    Agreed. If we’re dealing with labeling here in place of analysis, label that IRS jab as the fallacy of necessity.

  12. Steven Novellaon 02 Jun 2010 at 2:50 pm

    Sonic – it is not a non sequitur – saying that it has broad acceptance is just an argument from popularity. It is irrelevant and does not belong in a scientific paper. This paper is not a study of how acupuncture is used or accepted. The IRS statement is an absurd non sequitur – acceptance by the IRS is not even related to broad worldwide acceptance, it is a reflection of local politics.

    And you further missed my point about definition. While they may have given one definition of “acupuncture” in their article, they did not test one of the main components – “discrete points” – that was not tested as a variable. They also did not test whether needle insertion is a necessary variable either. So by their own definition – they did not test acupuncture.

    Their further discussion of acupuncture and its acceptance is not limited to their somewhat narrow definition, so they are doing exactly what I described – using research into a narrow definition of acupuncture (and not really doing that) to promote a more broad and vague definition of acupuncture.

    They did not even acknowledge that the definition they used is just one definition, that there are other ways to define it, and that acupuncture is used for many non-analgesic indications. Nor did they review the literature at all to give a sense that acupuncture in fact does not work for analgesia. Perhaps this is because their animal data does not extrapolate well to humans.

    In short – they totally dropped the ball. This was not a neutral scientific treatment of the topic. It was some interesting science wrapped up in bullshit.

  13. David Colquhounon 02 Jun 2010 at 3:18 pm

    I had a go at this one too, at http://www.dcscience.net/?p=3136. After the Nature news people got a bit upset about a comment I left on their blog, I thought I should explain.

    Essentially all of the newspaper reports misinterpreted the work. apparently because the press release from Nature Neuroscience misrepresented it too. But the press release just copied the tendentious spin in the paper itself. In the end, the people to blame were the authors themselves, and the reviewers for Nature Neuroscience who failed to spot the spin.

  14. Chad Brownon 02 Jun 2010 at 3:34 pm

    Great blog entry Doc. As a previous commenter already pointed out, you somehow manage to maintain a good amount of objectivity while still being outwardly skeptical of the topic in discussion. This makes your blog easier to read for someone who may actually be caught up in this bunk.

    Great write up. The End.

  15. Hubbubon 02 Jun 2010 at 4:34 pm

    Michael Meadon’s suggestion of writing to Nature Neuroscience reminded me of something I had been kicking around. I remember that you described the difficulty of getting the university to recognize the role of your blogging and internet contributions, as it all gets lumped under the umbrella of service.

    Perhaps, by taking your meticulously thought out criticisms to the editorials of the scientific literature, you may be able to bridge the academic world with that of your public outreach. Might such contributions make the educational and academic value of your blog more apparent as a serious medium for debating research (at least more serious than your typical run-of-the-mill blog)? Is it reasonable to think that such high-profile academic criticism might bring more academic attention to your web-based contributions?

    You also seemed to have made some substantive criticisms of the Cochrane Review on Homeopathy (a while back). Might you have a shot at the Bill Silverman prize for such a published critique?

    I hope that this suggestion is not viewed as presumptuous, trivial, or insulting. I know I have enough problems getting my own publications out the door, so I might roll my eyes at the suggestion that I just “publish more” (and I don’t contribute to four blogs and two podcasts and am not a clinician/professor/researcher/professional skeptic/father/birder). However, it seems you are already doing much of the work already in your spare time (not to diminish the time required to find the ideal references and formulate/edit your rhetoric carefully for a scientific publication). It would be inspiring to see your conviction to fight for high scientific standards pay off in academia, as opposed to largely being a trade-off with service.

    Best of luck!

  16. Steven Novellaon 02 Jun 2010 at 6:21 pm

    Hubbub – you are basically correct. The limiting factor is that it takes 10 times as much time to write something for publication than for a blog. But I do need to do that more.

    Recently, however, I am also working on yet another project, which will be done at the end of June. It’s almost impossible for me to do anything else until this is done.

  17. tmac57on 02 Jun 2010 at 6:42 pm

    I am in no way qualified to analyze a scientific paper, but as a layman, it looks to me like the authors might have been setting out to prove the mechanism for acupuncture, rather that to find a mechanism for local pain relief from local stimulation. Those are two very different propositions, and I don’t think the prominent acupuncture references reflected well on their credibility. I sensed a thread of bias throughout the entire study.But then, what do I know?

  18. ccbowerson 02 Jun 2010 at 9:30 pm

    “I am in no way qualified to analyze a scientific paper, but as a layman, it looks to me like the authors might have been setting out to prove the mechanism for acupuncture, rather that to find a mechanism for local pain relief from local stimulation.”

    What you have pointed out is a subtle distinction and is at the crux of the problems with this study. Your impression is exactly what I imagine the authors wanted to convey to reader, but there are several problems with this. First of all, in looking for a mechanism there is a huge assumption that there is a real effect of acupuncture (as they defined it- at discrete points). There is a lot of literature on the subject which does not support this assumption (comparing acupuncture to sham acupuncture for example). In addition their study did not even test for acupuncture as they defined it, so how could it evaluate for a possible mechanism? There was no proper control group for the acupuncture. The only clear conclusion that can be drawn involves possible “mechanism for local pain relief from local stimluation,’ which in this study used an acupuncture needle.

    Its unfortunate that an interesting study is contaminated by something that is either terrible writing + not recognizing the limitations of the study (at best), or bias that is meant to mislead (at worst).

  19. HHCon 02 Jun 2010 at 10:01 pm

    So according to Goldman et al, the inflicition of pain in alternative medicines, including acupuncture, chiropractic care, or massage therapy is responsible for the relief of pain. Will the best sadist win over his masochistic clients?

  20. Pinkyon 03 Jun 2010 at 12:31 am

    Since I was one (of probably a bazillion) to email this to SGU, I’ll post my main points:

    1. It’s a higher quality article relating to acupuncture than I’ve ever seen.
    2. It still has significant scientific reporting flaws which makes me question why it was published in a quality, peer-reviewed online publication.
    3. There is far too much bias (premise that acupuncture does work, and here we are trying to work out why…) and far too many generalisations in this paper for it to be of any clinical worth.
    4. I found it very alarming how poorly the research was presented on the actual academic research institution (Rochester) website.

    The NIH is intent on throwing money at these studies. They desperately need to increase scientific reporting quality which is the main thing lacking in this article.

  21. sonicon 03 Jun 2010 at 1:12 am

    I agree that using the IRS mention is bizarre and ‘fallacy of necessity’ or ‘argument from popularity’ might fit the bill-
    it was ‘argument from authority’ that I was objecting to.

    I agree they did not test acupuncture. But from the introduction-
    “Although the analgesic effect of acupuncture is well documented, little is understood about its biological basis.”

    The authors make it clear this is the assumption they are beginning with. They are not trying to test acupuncture (they assert it works), but are looking for a mechanism to explain the effect they consider ‘well documented’.
    The experimental evidence regarding acupuncture is such that there seems well read, knowledgeable people might disagree to what it shows.

    From a Bayesian perspective-
    1. What are the priors for a medical intervention that ‘balances the chi’?
    2. What are the priors for a medical intervention that ‘releases adenosine -an analgesic agent that suppresses pain through Gi-coupled A1-adenosine receptors’?

    I believe the priors are the reason for any disagreement as to what the evidence shows. Regardless, we must read the article allowing the assumptions. (Of course we can question the assumptions, but to test the logic we must allow them)

  22. veetvoojagigon 03 Jun 2010 at 4:33 am


    Excellent dose of clarity for this article. For what it’s worth, I also support the idea that you should write to the journal. I think that is especially true because of the high regard with which Nature Neuroscience is held. The correct forum would be in their ‘Correspondence’ section.


    I think this is the correct course of action when the peer review process fails to prevent such distortions. Authors – fail. Informal reviewers – fail. Formal reviewers – fail. Editors – fail. Diligent observers – win?

    Please write it, or if you do not have the time, encourage someone else to do so.

  23. BillyJoe7on 03 Jun 2010 at 6:51 am

    It is almost axiomatic:

    If you see a trial of acupuncture that seems to show that acupuncture works, look at the control they used. It is almost guaranteed to be inappropriate.

  24. Steven Novellaon 03 Jun 2010 at 8:21 am

    sonic – I think the main reason for the difference in interpretation of the literature by well-read academics is more than just prior assumptions – while this is likely the underlying cause.

    There is also a clear difference in process, that I have documented in numerous blog entries. The most egregious, off the top of my head, is the interpretation by well-read academics that the lack of a difference between “true” acupuncture and placebo acupuncture means that placebo acupuncture works too. Or that “pragmatic” studies are better than controlled trials.

    They have turned research logic on its head. It’s not subtle.

    For kicks, I have presented their arguments (without saying they relate to acupuncture, just “a treatment”) to colleagues and either their jaw hits the ground or they bust out laughing.

  25. Eric Thomsonon 03 Jun 2010 at 9:41 am

    Three points:

    1. Everything sonic said in the first three paras is exactly right. Given their assumption, the paper is solid.

    It’s like a paper that grants that pentobarbitol is an effective general anesthetic that proceeds to focus entirely on the mechanism of said effectiveness. There is absolutely nothing strange about this. Science builds on previous work all the time.

    2. So, given 1, is there a good analysis of the basic premise of their paper? That is, that acupuncture can relieve pain?

    I’m a little turned off by the nitpicking too. E.g., even if it does turns out that needle insertion and rotation can relieve pain, but it doesn’t depend on the location in a strong way, people get all bent out of shape like that is a falsification of acupuncture. It isnt’ clear at all this is true: it would just mean they have to revise their theory, not jettison the entire practice! It seems to me the needle insertion is pretty much the core methodology behind acupuncture, and all the stuff about chi and specific ‘centers’ is revisable within that methodological constraint.

    Some might say: “Yeah, sticking in needles may be effective to relieve pain, more than a sugar pill, but it isn’t as location-dependent as they say. You can stick it anywhere near the pain and it will work. Therefore, acupuncture is bullshit.” That to me is to admit acupuncture works, to grant a substantive defeat (for the acupuncture skeptic) coupled with a questionable and anemic semantic victory.

    So, given that’s how I’m looking at the semantic territory, is there a study that goes against what they say in the opening paras, and the citations therein? To be clear: does the methodology relieve pain more than a non-needle insertion placebo such as a sugar pill or sham insertion (they pretend to insert a needle but do something like give a little pinch to the skin)? I don’t care about the chi theory or the thesis that there are these special energy centers. I just want to know if the needle insertion works.

    And my third point gets directly at this question.

    3. People have been saying they didn’t do proper controls, but one nice internal control was the rotation of some of the needles, a common practice in acupuncture. When it was done, pain relief was increased significantly, as was ATP release, compared to when the needles were not rotated.

    Sure, that didn’t test the location-dependence of the effectiveness of acupuncture, but did get at the effectiveness of one common manipulation.

    People in this thread have been ignoring this internal control.

  26. ccbowerson 03 Jun 2010 at 10:13 am

    We are talking about a scientific paper and the complaints are not superficial at all given a general understanding of the overall literature on the subject. It is not just semantics. This doesnt mean that the study is not a good one in other ways. It is, but the false assumptions and the failure to put the study in proper perspecitve are very important. In fact there is a lot of spin in the article regarding acupuncture. This should not be overlooked as nitpicking.

    Also the problem is not that they didnt do proper controls. The design is fine, if they didnt try to use it as supporting a mechanism for how acupuncture works (as defined in their study). If you want to include everything that involves poking someone as acupuncture then you are going beyond what even the authors are stating (and against what people understand to be acupuncture).

    The confusion in these comments are just further evidence to me that looking critically at research is difficult, and can’t be done in a vacuum. How can we expect the public to be able to evaluate health information when it can mislead people who are towards the more intelligent end of the normal curve?

  27. Steven Novellaon 03 Jun 2010 at 10:34 am

    Eric – I did not ignore that control. I even provided a reference that bears on it – it is likely that rotating the needle increased insertional activity in the muscle which may increase adenosine release. This is just a way of increasing the stimulation.

    Also – as I pointed out – the evidence shows that poking the skin with toothpicks (or a sham needle) is as good as needle insertion, so the needle insertion part (not just the location) also does not seem to matter. That is why I conclude that acupuncture does not work – because BOTH core elements have been disproved.

    I also mentioned that other research into local tissue reactions indicate that any mechanical stimulation works – vibration, pinching etc, so it is odd that these researchers did not include such a control.

    This paper showed that adenosine release is part of a local tissue reaction to minor trauma and that binding to the A1 receptor inhibits pain. They did not show that acupuncture works.

    Further, it is deliberately naive in the extreme not to recognize explicitly that “acupuncture” is promoted and used for a host of non-pain indications. This mechanism can say nothing about that.

    Put into context, this paper feeds into a strategy that we see with many modalities – find some way to justify a very narrow effect from some non-specific aspect of an intervention, and use that to promote the broader intervention for a host of indications. It’s crank science.

    The take home point is that scientists need to be very careful to distinguish specific effects of an intervention from non-specific effects resulting from the process of the intervention. This study really only provides evidence for a non-specific effect, and promotes it as a specific effect. It’s sloppy, and the apologetics make it seem deliberately so.

  28. ccbowerson 03 Jun 2010 at 10:43 am

    Another thing that some people are missing is that this study has nothing to due with whether or not acupuncture is an effective treatment for pain in humans (or mice). There is already a lot of clinical research on the subject, discussed elsewhere. This is basic research that looks at a physiological mechanism for reducing pain activity in the peripheral nervous system in mice. This may not pan out to having a clinical benefit in mice, let alone humans if the clinical effect is small and transient. It may help us futher our understanding of pain in general, which I do not want to minimize.

  29. Eric Thomsonon 03 Jun 2010 at 10:58 am


    Just because other manipulations might produce the same effect doesn’t mean this one doesn’t. E.g., if you do a study that finds Pentobarbital works by activating GABA-A receptors, that doesn’t mean the study is invalid if there are other GABA-A R agonists out there and they didn’t compare pentobarbital to all of them.

    Besides, they explicitly discuss many of these issues in the paper (e.g., vibration on skin). Not including them seems a perfectly vanilla case of “beyond the scope of the paper.” They never said acupuncture is the only thing that does this, indeed they hypothesize that vibratory stimuli probably do the exact same thing. That is, the directly addressed this concern, and people acting otherwise is disingenuous.

    I apologize you are right I missed that you mentioned the twisting of the needles, I was focusing on the comments. People are failing to appreciate that it is a decent internal control, that demonstrates that this particular manipulation works, regardless of whether it is a special case of a more general category.

    Seems like a lot of straw men here, and people posturing as being more subtle thinkers than those who accept the paper. Having read the original paper I don’t buy it.

    A lot of people here obviously have prejudices against acupuncture that are polluting their reading of the facts. I know nothing about apupuncture other than it is sticking needles in people. Those of you who have been in the trenches, and associate acupuncture with all the specifics like chi and such, those of you who actually know a lot about the topic, seem to be reading way too much into this paper. If you read the original paper, they are just not promoting all that.

    The study was a good one for Nature Neuro, partly because it will be controversial. Obviously that is neither necessary nor sufficient for getting published there: the paper is otherwise solid, connects with things people are interested in, and will spur attempts at replication, criticism, etc.. I think the editors and reviewers made the right decision.

  30. skidooon 03 Jun 2010 at 1:56 pm

    Here’s to hoping “the peers” will overwhelmingly call-out this poo for what it is.

  31. Steven Novellaon 03 Jun 2010 at 2:20 pm

    Eric – you did not address my point that the paper confuses specific with non-specific effects. Your point about different drugs causing the same effect is a false analogy – that is not what I am saying.

    The point is if any mechanical stimulation causes the same effect then this is a response to mechanical stimulation – not acupuncture. Again – the research shows, you don’t have to stick needles.

    You seem to want to take this study out of context. But ignoring that context is dangerous, and good scientific papers are supposed to put themselves into proper context. These authors put their paper into a false context, and made many assertions that are either untrue or fallacious. That is what I was complaining about.

    The actual research is fine, as far as it goes.

  32. ccbowerson 03 Jun 2010 at 2:46 pm

    “Just because other manipulations might produce the same effect doesn’t mean this one doesn’t.”

    It does if those other manipulations are used as controls in other studies to see if acupuncture works. If something does not outperform a control, you can’t conclude that it works. Acupuncture has failed in this regard in many clinical trials in treating pain. Your argument is that both acupuncture and the controls in these studies work? Its not logical. Lets say that you have a drug trial in which drug X did not outperform placebo… but both produced some effect on variable Y. Would you conclude that placebo works too?? Of course not.

    “E.g., if you do a study that finds Pentobarbital works by activating GABA-A receptors, that doesn’t mean the study is invalid if there are other GABA-A R agonists out there and they didn’t compare pentobarbital to all of them.”

    This analogy is not apt. Phenobarbital has known effects that have been demonstrated in previous studies. The GABA-A analogy does not make sense here, since in your example you are looking for a mechanism of a known effect. You keep forgetting that there is already research that shows that acupuncture does not work for the treatment of pain when compared to a proper control. That does not mean that it has no physiological effects. Its just that those effects are not unique to acupuncture.

    “A lot of people here obviously have prejudices against acupuncture that are polluting their reading of the facts. I know nothing about apupuncture other than it is sticking needles in people.”

    So your argument is that having prior knowledge that informs and puts things in perspective is a bias? If we all knew nothing about a subject we would be better able to judge it? Prior knowledge is prejudice? Come on. That is absurd. How can you properly evaluate a study with little or no knowledge of the subject? You can’t. That is why you are struggling here.

  33. Eric Thomsonon 03 Jun 2010 at 3:39 pm

    Steven Novella said:
    Again – the research shows, you don’t have to stick needles.

    My claim was that they are well aware that they may be studying a species of a genus, which they discussed explicitly on page 5 of the paper. That doesn’t invalidate the work, or the efficacy of one of those species.

    They know there are other ways to produce the same effects. They did it themselves, by injecting A1-R agonists after all! They never claim that nothing else can have these effects, that acupuncture is unique in this regard.

    That doesn’t mean this particular species of manipulation doesn’t work, or that it is identical to the other species.

    ccbowers: I knew someone would say that, I opened myself up to that last criticism :)

    My point is you guys have too much experience with kooky acupuncture propaganda you have studied, and that is distorting your vision here. This paper clearly isn’t kooky acupuncture propaganda, it is a good study. There was a cheeky paragraph in the introduction, I admit.

    On the general point that they have assumed that it works better than placebo, that is a good one that I’ve admitted multiple times, and requested references on that particular topic (e.g., a counter to reference four in their introduction).

    So in response to:
    You keep forgetting that there is already research that shows that acupuncture does not work for the treatment of pain when compared to a proper control.

    I want to see the studies.

    On the other hand, their internal control mentioned above seems to be very good evidence that acupuncture (in the way they used it) worked for suppressing pain, significantly more than simply inserting a needle. Novella has admitted as much, he’s just trying to say to say this isn’t significant, and doesn’t actually support acupuncture because it is a species of a genus.

    My claim is, fine! Biology studies species all the time! We study pentobarbitol within the genus of anesthetics, etc etc.. now that they’ve done this, they can look at A1-R distribution in the peripheral nervous system, find out how they align with acupuncture points, etc..

    But on the original point that they simply assume it has analgesic qualities, despite what I said this isn’t exactly right, given their internal control. But in support of their claim, they cite:
    Zhao, Z.Q. Neural mechanism underlying acupuncture analgesia. Prog. Neurobiol. 85, 355–375 (2008).
    Which isn’t a great paper to cite, for one it is sort of kooky. For another, it doesn’t actually address that question but is more like this paper.

    One main paper I found critical of acupuncture was:
    H.H. Moffet, Traditional acupuncture theories yield null outcomes: a systematic review of clinical trials, J Clin Epidemiol 61 (2008), pp. 741–747.
    Despite the title, the data are seem much less clear. For instance, he says in the abstract “Clinical trials demonstrate that acupuncture can affect outcomes and is distinguishable from a placebo.” His point was that the theories behind acupuncture don’t actually seem to help produce such effects. Note also he didn’t specifically narrow down his research to analgesic effects.

    In other words, acupuncture as defined in the paper under discussion seems to have real effects. Their internal control supports this. I’m very curious about studies suggesting it does not (and note the paper should have proper placebo controls: if the “control” is a comparison to somatosensory stimulation that may induce ATP release and analgesia, that would not be a good control: it would be like saying something doesn’t have analgesia effects because it didn’t do better than aspirin!).

    Also check out ‘Reduction of chronic non-specific low back pain: A randomised controlled clinical trial on acupuncture and baclofen.’

    The more I look into this, the more it seems ya’ll are wrong, at least as far as universal dismissiveness goes.

    We should be happy they are actually applying the scientific method to this technique.

  34. ccbowerson 03 Jun 2010 at 4:42 pm

    “This paper clearly isn’t kooky acupuncture propaganda, it is a good study.”

    It is a little bit of both, and that is what is causing the harsh-ish criticism. It feels more frustrating to me to see good research tainted than seeing bad research, since the potential was there. The way a study is framed and interpreted in the study is nearly as important as its methodology.

    Regarding the evidence and acupuncture: Its cloudy because it is a subject for which there are a fairly large number of trials, but the quality of those trials varies widely. Someone could potentially come up with studies to support an erroneous point of view, but those studies would be of poor quality.

    “On the other hand, their internal control mentioned above seems to be very good evidence that acupuncture (in the way they used it) worked for suppressing pain, significantly more than simply inserting a needle.”

    The caution that I have to say about this is that we already have human clinical trials on this topic, so we can’t extrapolate the basic research in animals to clinical effects in humans in a way that supercedes the research in humans. Thats a bit backwards.

    I would like to look up some trials for you, but I have to get back to work.

  35. Robbon 03 Jun 2010 at 4:48 pm

    There was a request for the evidence that contradicts the introductory sentences. This is an example, from the Acupuncture Wikipedia page, which I help edit:

    “An analysis of 13 studies of pain treatment with acupuncture, published in January 2009 in the journal BMJ, concluded there was little difference in the effect of real, sham and no acupuncture.”

    Madsen, MV; Gøtzsche, PC; Hróbjartsson, A (2009). “Acupuncture treatment for pain: systematic review of randomised clinical trials with acupuncture, placebo acupuncture, and no acupuncture groups”. BMJ 338: a3115. doi:10.1136/bmj.a3115. PMID 19174438.

  36. Eric Thomsonon 03 Jun 2010 at 5:52 pm

    Many thanks Robb I’ll look into it, and also I found the NIH site on the topic pretty helpful. The following is good because it breaks things up by type of pain:
    Kelly, RB (2009) Acupuncture for Pain. American Family Physician. 80: 481-484 (link here.

    These bits convince me that data in humans is somewhat murky and inconclusive and depends strongly on the type of pain being addressed. The NIH consensus in 1998 published in JAMA was:
    “Acupuncture as a therapeutic intervention is widely practiced in the United States. Although there have been many studies of its potential usefulness, many of these studies provide equivocal results because of design, sample size, and other factors. The issue is further complicated by inherent difficulties in the use of appropriate controls, such as placebos and sham acupuncture groups. However, promising results have emerged, for example, showing efficacy of acupuncture in adult postoperative and chemotherapy nausea and vomiting and in postoperative dental pain. There are other situations, such as addiction, stroke rehabilitation, headache, menstrual cramps, tennis elbow, fibromyalgia, myofascial pain, osteoarthritis, low back pain, carpal tunnel syndrome, and asthma, in which acupuncture may be useful as an adjunct treatment or an acceptable alternative or be included in a comprehensive management program. Further research is likely to uncover additional areas where acupuncture interventions will be useful.”

    Getting back to this study: I can’t avoid the strength of the internal control in this study. The fact that it is in rodents makes me more, not less, likely to believe it. No chance for a placebo effect. Also, think about it logically: injecting an A1-R agonist relieves pain (If A1-R agonist, then less pain). Acupuncture releases an A1-R agonist (data from study). Apply modus ponens.

    Thanks for the reference Robb.

  37. Robbon 04 Jun 2010 at 5:34 am

    @Eric, I dunno man, the NCCAM is sketchy: Web pages and documentation on their site is absolutely no substitute for solid research. The site’s 80% PR and 20% science. Have you thought that you’re possibly making an appeal to authority by calling it “an NIH site”, instead of NCCAM?

    I read the abstract for the Kelly article (no methodology discussed), and the consensus as well. The consensus sounds like it was solid work, but … with so many negative results out there … especially as the studies are higher quality and more modern … this field of research is simply … uninteresting.

    And for sure not worth torturing animals for. If we’re going to do that (inflict pain and tissue damage to measure the response) let’s *only* do it for genuinely valuable, promising areas of research.

  38. Eric Thomsonon 04 Jun 2010 at 10:58 am

    The Madsen paper clearly notes acupuncture is helpful; the NIH consensus is clear that in some cases it is beneficial, but in other cases it is very murky. The internal control in the study we are discussing makes the case very strong, and nobody has a good answer to that.

    This is clearly a promising area of research. If the results were guaranteed beforehand we wouldn’t need to do the science. Isn’t that sort of the point of scientific study of alternative medicine? Do we really expect all alternative methods to turn out to have nothing of value? It’s an empirical question, and we need to follow the data.

    As for negative results, the fact that the controls in many of these studies may induce local ATP release makes it even harder to interpret in those instances when there isn’t a significant difference between acupuncture and control (like using aspirin as a control in a study of analgesia). This is itself an empirical question.

    Also, these meta-studies are sort of watered-down science in my opinion. I would rather pick one well-designed study and examine it in detail than trust someone drunk on pubmed to do the metaanalysis, lumping together studies with very different questions (“pain” is an extremely heterogenous category). That said, if you look at the plots in Madsen’s paper what jumps out at me is the number of studies where the mean is to the left of the ‘no difference’ line.

    In many of the critiques of acupuncture I’ve been reading the past couple of days, one of the criticisms is there is no “known mechanism.” Now they provide one, undermining a major objection. That said, I always thought that was a stupid objection. We can observe X without knowing the underlying mechanism, that’s the starting point for looking for an underlying mechanism!

    I was thinking a lot last night about the controls, how hard it is to do controlled experiments with acupuncture. This is the most serious problem, in my opinion. It is nearly impossible to have the research properly blinded. Same goes with physical therapy, the best they do with PT is controlled trials, and there are a lot more of those coming out with acupuncture, and despite the spin I’ve seen here, many of those are giving positive results.

    The bigger question is whether the benefits outweigh the costs. When it comes to pain management, they might.

  39. Eric Thomsonon 04 Jun 2010 at 11:04 am

    On that NCCAM site being not objective, they cite everything you cited and more, and make most of the same points. It is a good site.

  40. Eric Thomsonon 04 Jun 2010 at 11:09 am

    Indeed, reading it over more closely the NCCAM site makes me more skeptical of acupuncture’s efficacy than the people here! It seems very hard to dpull out the placebo effects. What if they tried acupuncture on someone under general anesthesia to get rid of placebo?

  41. Robbon 04 Jun 2010 at 5:09 pm

    @Eric, I think we need to back up. As you wrote, “I know nothing about acupuncture or studies of acupuncture.” You should read up on it:


    I say this because the biggest problem (really, the only problem) with this study is that it has nothing to do with acupuncture. And, you use the word, “acupuncture”, in your posts here, so we need to clarify this.

  42. tmac57on 04 Jun 2010 at 9:49 pm

    Eric, look at it this way: If there were some study that tentatively found that hearing sound stimuli, while in a coma, helped the recovery of a patient in some measurable way, would you feel okay about the authors repeatedly implying the benefits of audible prayers ,instead of the more generic audible stimuli? That is why the specificity of terminology matters here. There is a long history of cultural and popular assumptions about both acupuncture, and prayer, that influence the readers of such a study,so that they will be swayed by their biases,not only toward, but also AGAINST the conclusions of such a study. You see, the authors may have unwittingly caused some unwarranted bias that is antagonistic to their core finding, by the heavy use of the acupuncture model.

  43. Eric Thomsonon 04 Jun 2010 at 10:55 pm

    Robb: I am somewhat familiar with the weird philosopy behind acupuncture. My point above was that there’s different levels of commitment to that theory, from the hard core chi-wu-li master types, down to the people who just want to study the effects of sticking acupuncture needles in people, regardless of underlying weird background philosophy.

    This study seems in the latter category, not the former, but my objection to many of the comments here is that they are responding as if it were the former. So, yes that’s acupuncture, but so is this which uses a more relaxed and standard parlance among the less embroiled in these debates.

    Put it this way, will grandma give a crap if it is shown that sticking needles in helps her pain, but that the location isn’t important like the old chinese herbalist claims? No. But she’ll still say it’s acupuncture.

    tmac: I like that analogy, it is clever, but I’m not convinced it is apt. The question is whether acupuncture is really a species of analgesic or a species of analgesic plus a bunch of other crap that isn’t doing anything E.g., pentobarbitol plus a waving of a wand above the patient for general anesthesia.

    I don’t see the wand waving here, frankly, just a twisting of the needles that had a measurable effect. If they had gone on about chi, about the metaphysical crap, then it would be analgesic plus wand. Not in this study I don’t think.

    That said, that NIH site and sources therein has convinced me that the results are extremely hard to interpret, placebo effect for some reason is huge. I’m turning sort of agnostic on the whole acupuncture thing, whereas yesterday I was leaning toward theism.

  44. ccbowerson 04 Jun 2010 at 11:44 pm

    Its funny that a person who admits to having little to no a priori knowledge of acupuncture has been defending it consistently based upon such little evidence/knowledge. Note that nearly all the people you have spoken to (including the writer of this blog) are pretty much saying the same thing. I’m not trying to bring up an argument from authority as much as I would have expected a little less jumping to conclusions that are counter to what people are telling you. I realize you were, the entire time, open to being convinced otherwise, but you were coming at this from the wrong angle.

    I admit the evidence is clouded by a lot of bad studies, but the best studies point to a small transient effect that is nonspecific, and does not appear to hold up when compared to certain controls. Again, the study discussed is not evidence that acupuncture works, we have clinical data in humans that specifically test the efficacy of acupuncture.

    “That said, that NIH site and sources there in has convinced me that the results are extremely hard to interpret, placebo effect for some reason is huge. I’m turning sort of agnostic on the whole acupuncture thing, whereas yesterday I was leaning toward theism.”

    You are moving in the right direction here, but… I second the notion that the NIH/NCCAM website is extremely soft regarding acupuncture. This is due to the nature of the NCCAM – they have an inherent bias of keeping themselves relevant despite the few positive well done studies. They have obvious biases in their presentation of CAM. Here is something Steve wrote in science based medicine:


  45. Eric Thomsonon 05 Jun 2010 at 12:37 am

    Folks here are doing a pretty poor job of handling someone playing devil’s advocate here. If it were like creationism, you would hopefully have much better arguments against me. The studies you have cited (well, make that singular, the one study) explicitly said that acupuncture works. OK, the effects are small and transient. So is aspirin.

    I’m glad you aren’t going to try to pull any appeal to authority, because for one it is clear the true authorities to be found in academic journals are divided on acupuncture. For another you should be able to make the case on the basis of good arguments and evidence. Don’t worry, I can handle complex thinking, primary research articles, statistics, medical jargon, am very familiar with the logical fallacies, etc..

    The reaction to the NIH acupuncture page seems generic and ill-defined. What specifically is biased about it? These general ad hominem arguments against the site don’t cut it. I read the site, I have a perfectly sound critical mind, and ability to evaluate literature and literature reviews, and they do a very good job of summarizing the weaknesses of the case for acupuncture, much better than anyone here!

    Even Dr Novella in his criticism that you cite says “I will acknowledge that there can be some reasonable disagreement about whether or not acupuncture is useful for some symptomatic treatment, like pain.” It’s a good study. Stop reading chi and meridians and kooky eastern metaphysics into it. Read the bloody paper from front to back. Aside from a cheeky paragraph in the intro (which everyone knows is not where the science is), point out a specific problem with the paper. The addressed Novella’s concerns about potential nonspecificity, for instance, so that’s a non-starter.

    Enough with the patronizing, I overstated my naivete about acupuncture, you won’t be able to intimidate me into agreeing. If you want to convince me use evidence, arguments, reason. I’ve done a literature review over the past two days, and it is clear that things are not settled. It depends on the treatment target, etc, and we need more controlled trials.

    Hell, try reading the NCAAM site it seems you guys need to learn more about the studies that undermine acupuncture. :)

  46. BillyJoe7on 05 Jun 2010 at 2:15 am

    Steven Novella,

    “The most egregious, off the top of my head, is the interpretation by well-read academics that the lack of a difference between “true” acupuncture and placebo acupuncture means that placebo acupuncture works too.”


    I remember a systematic review of acupuncture in the prophylaxis of migraines which concluded that “true acupuncture works as well as sham acupuncture” instead of “acupuncture works no better than placebo”.

    It was even suggested that having proven the value of acupuncture in migraine prevention, we could now dispense with sham acupuncture and concentrate on fine tuning the protocol!

    Worse still, this was a systematic review in the Cochrane Database,

  47. BillyJoe7on 05 Jun 2010 at 2:53 am


    It historical discription of acupuncture looks like this:

    Stage 1: Qi, Meridians, Sticking needles in acupuncture points.
    Stage 2: Meridians, Sticking needles in acupuncture points.
    Stage 3: Sticking needles in acupuncture points.
    Stage 4: Sticking needles in anywhere.
    Stage 5: Not sticking the needles in.

    When do you stop calling it acupuncture?

  48. ccbowerson 05 Jun 2010 at 11:35 am

    “Aside from a cheeky paragraph in the intro (which everyone knows is not where the science is), point out a specific problem with the paper.”

    Its not just the intro, but thats a large part of it. You don’t see the problems with misrepresenting your research, even when the methodology is sound? Not only is the study not in the proper context, it is the wrong conext. The actual findings of the study are very interesting, but that doesn’t excuse the obviously biased perspective.

    Also, the implications that the study support that acupuncture works is a misread of the study. No one here will agree with that (again its an animal model that isn’t testing efficacy in humans). I can come up with a study that shows that a drug interacts with certain receptors, but until I show that acutal humans taking the drug for a specific condition, I cannot say it works. In fact there are thousands of drugs that “should work” but don’t in real world applications.

    “It historical discription of acupuncture looks like this:

    Stage 1: Qi, Meridians, Sticking needles in acupuncture points.
    Stage 2: Meridians, Sticking needles in acupuncture points.
    Stage 3: Sticking needles in acupuncture points.
    Stage 4: Sticking needles in anywhere.
    Stage 5: Not sticking the needles in.

    When do you stop calling it acupuncture?”

    This is a good assessment of why the term acupuncture is problematic. The definition is becoming more broad inorder to support a more narrow definition.

    ** Heres a somewhat hypothetical analogy that may help: Lets say there is some condition X for which a chiropractic manipulation Y seems to help when compared to doing nothing. Lets say it is then tested against chiropractic manipulations A, B, and C and they all work versus doing nothing, but none stand out as better than the others. Lets say that manipulation Y is again tested against a massage by a therapist for condition X, and again they perform better than nothing but neither is better than eachother. Then lets say manipulation Y is compared to random touching in the general area by an untrained person and you get the same results – better than nothing, but one is not better than the other.

    Do you conclude that chiropractic works for condition X? I don’t. If you do, then that is the difference in our discussion. Now I can’t say that this is true for every single condition for acupuncture, but much of the research for pain points in this direction. The data are unimpressive and even where there may be something there (for the ‘genus’ as you say) its not a robust effect.

  49. Eric Thomsonon 05 Jun 2010 at 6:02 pm

    ccbowers, inapt analogy. They compared X and Y and found X significantly better than Y, not that neither is better than the other. Again, read the paper and find a criticism other than the first paragraph, where the science isn’t.

    On the definition of acupuncture, that is going to be somewhat arbitrary. I would say up to 3 it is clear, and whether 4 is counted is a matter of some convention at this point.

    What is wrong with them allowing for revision of the locations based on updated evidence, so the ancient points are revisable? For instance, what if they find a location with a huge density of A1-receptors using a mouse model, then try the analogous location in humans, and it turns out to have really strong analgesic effects in accord with the predictions of this experiment? Would that be acupuncture? Yes, especially based on the way they used the term in the study.

    A field of inquiry can undergo radical transformation. E.g., astronomy went from weird geocentric firmament based theories to what we have now. It’s all still astronomy. My hope is that will happen with acupuncture. That’s not to say a lot of the practice of acupuncture won’t turn out to be mostly bullshit with a little ATP release thrown in for good measure, but eventually the science will sort that out. It is quite clear that it hasn’t been sorted out yet.

    OTOH, if it turns out to be no stronger than aspirin or local adenosine injection, then perhaps we shouldn’t have insurance cover it. There’s the clinical cost-benefit analysis side, and the science side. The basic science is clearly far from settled. The clinical stuff is also not settled, but if I had to make a prescription I’d go with aspirin or steroid injections or physical therapy at this point in time.

  50. ccbowerson 05 Jun 2010 at 6:41 pm

    My analogy was clearly addressing the evidence for acupuncture in general, and not his particular study, which did not address the efficacy of acupuncture. I think it is a apt analogy. Listen to Steve’s comments this week on the SGU… He is spot on. Perhaps verbal is easier than written words since it is hard to get everything across in this setting.

  51. Eric Thomsonon 05 Jun 2010 at 6:50 pm

    But the comparison in this study does get to the efficacy of acupuncture.

    Anyhoo I may take a listen if you truly think he says anything that hasn’t already been said here.

  52. ccbowerson 05 Jun 2010 at 9:32 pm

    “But the comparison in this study does get to the efficacy of acupuncture”

    I think you may get something out of Steve’s take in this week’s SGU because he points out the reasons why this study does not really address the efficacy of acupuncture, wording things in a way that may be more effective than my comments.

    Keep in mind that in this entire discussion I (and I think others) am not really biased against this study because of acupuncture, but I am trying to frame the results in light of what I know about the subject from other studies I’ve read about. This is already a well studied subject.

    In this week’s SGU, Steve does a good overview of the study, the possible implications of the results, and addresses the possible limitations of the rodent model in extrapolating to humans. He mentions a few things he doesn’t mention here.

    Its not that I am denying that there is a physiological effect going on here in this adenosine study…there definitely appears to be. But in order to reconcile this with other studies I am concluding that the effect is small, transient, and localized enough so that it is not clinically relevant in the treatment of many types of pain (the way that it is currently utilized). Thats not to say that nothing will come of the results of this study. I hope that something does come of it, but it doesn’t change my overall impression of acupuncture itself.

    I think these exchanges have at least progressed somewhat, and I can’t say that of all exchanges that I’ve had on here. If nothing else I’ve put things into better perspective in my own mind.

  53. dwhiteon 12 Jun 2010 at 4:46 am

    I’ve read through the comments and original article but have not at this time read through all the studies cited in these comments.

    Further disclaimer:
    I am presently studying Traditional Chinese Medicine (TCM, including acupuncture). My undergraduate degree is in Physics.

    There is a very large issue that no one here has mentioned and I’d like to address it. TCM is a 3000+ year old tradition that originates in a completely different culture and language than the one that trained us. It has a completely different paradigm than western medicine (which for the purposes of this comment I’ll refer to as biomedicine).

    It’s origin in a completely different language creates significant translational issues. Perhaps the most significant is that a single word in Chinese frequently can be translated into multiple different concepts in English, only some of which can be clarified by contextual clues. This creates a serious obstacle to understanding between TCM and biomedicine. It’s difficult for me to convey in a short comment how serious an issue this is but I’ll try to give you a simple example.

    Imagine an experiment to try to determine whether acupuncture can significantly treat knee pain. From a biomedical paradigm, as long as a test subject’s pain is in the knee, he is a valid test subject. From a TCM paradigm, knee pain is merely a symptom of some underlying condition. There are numerous known conditions in TCM that cause knee pain, but each one can have a different treatment protocol which depends on the cause of the knee pain. Further TCM, is extremely patient specific. Even if two patient suffer from the same underlying condition, it’s possible that other complications would change the optimum treatment, although for research purposes this is less of an issue. You should also be aware that acupuncture treatments rarely involve a single needle (although there are techniques that do).
    Please understand, when you see research on acupuncture that doesn’t include a doctor trained in TCM you’re probably going to get poorer results because the researchers don’t understand how to design the experiment in a way that recognizes how acupuncture treatments work.

    Another issue is so-called “sham acupuncture”. This concept is essentially flawed. Acupuncture points aren’t points, they are areas. The size of the area varies by both the particular point (some are relatively small, others are relatively large, there are some as large as a quarter) and to a lesser extent by the patient’s physiology. The location of these points is based on the physiology of the patient and can be fairly difficult to locate accurately. There are over 200 known acupuncture points, so it can be surprisingly difficult to stick a needle and not hit one. It’s also possible to apply a needle to an area that would aggravate a condition instead of helping. So attempting to use a needle inserted randomly in not really a good control. You should also be aware that insertion of a needle isn’t necessarily required in the practice of acupuncture, so pinching, poking with toothpicks and other such methods do not constitute a good control either.
    It’s probably going to be more effective in research to treat acupuncture more like research for new types of surgery. Generally, cutting a patient at random isn’t seen as a good control for surgery.

    I realize that this makes designing good research for acupuncture very challenging but that shouldn’t stop us from doing the research. The studies that I have seen, both in support of but more commonly not in support of acupuncture have often been deeply flawed from a TCM perspective so any conclusions presently drawn from the research literature to date is probably premature.

    Food for thought, gentlemen.

  54. Steven Novellaon 12 Jun 2010 at 11:23 pm

    dwhite – forgive me but that is just a bunch of excuses for negative research. You can transpose what you said over the same excuses for ESP, free energy, or what have you. When people start throwing around the world “paradigm” as an excuse for negative research, that’s a huge red flag.

    Also – the concepts behind acupuncture as used today are not that ancient and not uniquely Chinese. Here is a good summary of the history. http://www.sciencebasedmedicine.org/?p=583

    So – if you can poke some toothpicks anywhere and basically get the same effect, that actually is quite a good control. It means there is nothing to all the specific nonsense of acupuncture – acupuncture points or regions or whatever are just fairytales. There are just non-specific effects and placebo effects.

    There is also no mechanism or plausibility to any specific effects.

    Acupuncture is superstition. It is useless as a concept – it contains no testable demonstrated unique principles.

  55. Steven Novellaon 12 Jun 2010 at 11:27 pm

    Eric – how do you think the modern acupuncture points came about? Extensive testing, even trial and error? That is not what history tells us. At first they were over blood vessels, used for lancing. Recently they were moved over nerves because the philosophy changed specifically to distance the practice from blood letting. It was political – not scientific.

    There is no science, either basic or empirical, to acupuncture points.

  56. ccbowerson 13 Jun 2010 at 12:11 am

    “So attempting to use a needle inserted randomly in not really a good control.”

    Why is it not? There is no argument made. Because random needle sticks may actually conform to acupuncture points? If there are so many points that random sticks in various places works, then what is the utility of selecting specific points. Right, there is none.

    “forgive me but that is just a bunch of excuses for negative research.”

    He is very close to saying “western medicine” can’t detect the effects of “eastern medicine” because of a language barrier.

  57. neurosciguyon 27 Jul 2010 at 9:01 am

    I’m very late to this post, but I wanted to point out the key control that the authors of the paper didn’t perform. The mouse “acupuncture” effect was achieved by inserting the needle and then twisting the needle every 5 minutes for a 30 minute duration. Any knowledgeable scientist working with mice, especially a behavioralist or pain researcher, knows about stress induced analgesia. Simply stressing the mouse (or rat), as would have absolutely occurred in this instance, imparts a HUGE analgesic effect (a enormous problem for true pain researches to control for in their experiments). The PROPER CONTROL would have been having a separate cohort of mice where you took them out of the cage every 5 minutes over a 30 minute course, just like the “acupuncture” group, but NOT PERFORM THE ACUPUNCTURE. It is amazing that the reviewers of the paper allowed this to slide since it COMPLETELY EXPLAINS THE ACUPUNCTURE effect. What we don’t know is the editorial slant here. Did they not send this out for review to a pain researcher? If not, that is an egregious error. What’s interesting is that this stress induced analgesia IS dependent on this particular adenosine receptor. It’s a shame that the real science of the paper is clouded by the acupuncture voodoo.

  58. Eric Thomsonon 27 Jul 2010 at 10:39 am

    Steven asked:
    how do you think the modern acupuncture points came about? Extensive testing, even trial and error? That is not what history tells us.

    That’s an interesting point. Perhaps they will be able to find better points using this kind of research, e.g., the locations of maximum A1-receptor density in the body. But what if it turns out that the most popular and beneficial (according to practitioners) points ended up coinciding with such locations? Would you still say it’s bull?

    You make an excellent point (people interested could check out Kelly, D., ed. Stress-induced analgesia. New York: New York Academy of Sciences; 1986. (Ann. NY Acad. Sci., vol. 467).

    However, I think we can cut out the alternative stress-based explanation a couple of ways. Most directly, if it were a generalize stress-induced response, the analgesia effects would be seen even if the acupuncture and twisting were done anywhere on the animal, and they showed (Supplementary Figure 3) that when they injured one leg, and did acupuncture with twisting on the other leg, the analgesic effects were not observed.

    Second, the fact that they were able to kill pain pharmacologically (and take away such pain reduction in KO mice) in ways predicted by the observed effects of the needles suggests it wasn’t a general stress response.


    More generally, now that I’ve had time to think about this thread and let it simmer, I still largely agree with my assessment. Yes, Steven may be right that you could get similar effects using different techniques, but that doesn’t mean they didn’t get the effects with this technique (one person studies Tylenol, another aspirin but they both relieve pain). The authors even discussed this possibility in the final bit of the paper, so it just seems a sort of banal point. People here, I still believe, overreacted to a bit of cheeky stuff in the introduction, but the study overall is solid and pushes research into acupuncture in interesting directions as it suggests predictions about localization of needles based on A1-receptor densities.

  59. neurosciguyon 27 Jul 2010 at 11:40 am

    Eric Thompson,

    They state right in the figure legend to Supplementary Figure 3 that “The acupuncture needle was inserted in the Zusanli point and remained for 30 min, but not intermittently rotated every 5 min”. Thus, the animals were not stressed every 5 min for 30 min as in the “acupuncture” group. Now that I think about this, I think the AUTHORS ARE INTENTIONALLY MISLEADING with this experiment. Why would you go to the trouble of doing this “acupuncture” control, but not go all the way and take the mice out every 5 min to handle them…probably because they knew it would entirely explain their effect. The pharmacology is a moot point because I’ve already said I believe they have a real pharmacological effect but this says nothing about how “acupuncture” would work.

  60. Eric Thomsonon 27 Jul 2010 at 11:57 am

    Sorry I meant Sup Figure 1, which goes with “normal” figure 3, which for some reason doesn’t include the contralateral leg on the same graph. Not sure why they relegated that to Sup Fig 1 (They say in Fig 3 legend that “The sensitivity of the contralateral (control) leg to mechanical and thermal stimulation from these experiments is shown in Supplementary Figure 1.”).

    Figure 2 also includes some relevant data. However, you are right about Sup Figure 3, it’s a bit strange.

    At any rate, it seems they did do the proper controls with the contralateral leg. The way the present the data makes it harder to interpret though, and your general caveat is a great one to consider for all these acupuncture studies.

Trackback URI | Comments RSS

Leave a Reply

You must be logged in to post a comment.